America’s Most Infamous Nuclear Site Returns to Fuel the AI Boom

0:00 So lots of detail in here about the rebirth to nuclear.

0:02 I wanna start with just what happened in 1979

0:05 and and that nuclear disaster here in The US.

0:08 I was, surprised to learn reading your story

0:11 that even people who had been exposed

0:14 in the area to what was released as a result of that, according to a study,

0:19 really only received as much radiation as would would happen in a chest X-ray.

0:23 Less, actually Less.

0:24 Than a chest X-ray.

0:26 Less.

0:26 Yeah.

0:27 It I mean, obviously,

0:28 it was a huge public relations disaster for the nuclear industry,

0:33 and that led to all kinds of heightened oversight,

0:37 new safety policies, lots of scrutiny.

0:41 It really slowed down the nuclear industry in The US,

0:45 and it was justifiably a disaster.

0:47 People were terrified.

0:48 Thousands and thousands of people fled the area.

0:51 But when they eventually, you know, went and studied it and measured it,

0:56 the amount of radiation that was released was not deadly.

0:59 It was not huge.

1:01 Mhmm.

1:01 But just the idea of radiation is terrifying everybody.

1:05 You can't see it.

1:06 You can't smell it.

1:07 You can't feel it, then you don't know what's happening.

1:10 It's scary.

1:11 And yet here we are in 2026, Will.

1:13 And I feel like in the last couple of years, we've done nothing.

1:16 I think about that NJIT event that we did,

1:19 and we did a panel of a stage and what they wanted to talk about.

1:22 '22, 2023?

1:23 It's a few years.

1:24 Yeah.

1:24 Nuclear.

1:24 And I think it caught us by surprise.

1:26 You've been talking about it and reporting it out here.

1:29 Is the whole difference just the big tech community,

1:32 the AI craze, and the demand for power?

1:34 Yeah.

1:35 Pretty much.

1:35 I mean, for a while,

1:37 nuclear was really kinda waning, but mostly because it's expensive.

1:41 It you know, these plants are big.

1:43 They're complicated.

1:43 They take a long time to build.

1:46 Like I just said, there's tons of scrutiny.

1:48 So they're expensive to build and operate.

1:50 So until about, like, 2019, 2020, 2022,

1:55 they were just shutting down for economic reasons.

2:00 They were just more expensive.

2:01 You could get cheap natural gas.

2:02 You get cheap wind and solar.

2:04 But now we need electricity, and it is.

2:08 It's big tech.

2:09 It's AI.

2:10 They need these data centers just need so much electricity,

2:13 and they want it as fast as they can get it,

2:17 and they want as much as they can get.

2:19 And nuclear can deliver a lot of power.

2:21 Well, you point out in the piece, this is not just about The United States.

2:24 I mean, there are countries all over the world

2:26 that moved away from from nuclear in Europe,

2:28 Japan after Fukushima, and now they've all changed their minds too.

2:33 They haven't all changed their minds.

2:34 Germany is Many of them have changed their minds.

2:36 Excuse me.

2:36 Thank you.

2:37 I've been tracking this for a long time.

2:39 There's been a huge reversal, and people really are starting to embrace nuclear.

2:43 It's I mean, not everyone in the world,

2:46 but it's really getting a lot of acceptance.

2:49 And I I should point out, I mean, there's 3 Mile Island.

2:53 There's Chernobyl.

2:54 There's Fukushima.

2:55 Mhmm.

2:55 That oh, that's it.

2:57 There have been three major accidents.

3:00 How does that compare to But but Chernobyl was really bad.

3:04 Chernobyl was really bad.

3:06 I wanna downplay.

3:06 It was really was really bad.

3:08 I mean, was really bad, but nobody died from radiation.

3:12 They died from the flooding.

3:13 But Chernobyl, like, we've seen, you know, that whole area, just nothing.

3:19 Because, you know, so when it goes bad, it goes bad.

3:22 When it goes bad, it goes bad big.

3:24 Yeah.

3:24 Which is why it is so heavily regulated.

3:27 And I we looked at some death statistics because I mean, let's be real.

3:31 Living on this planet will kill you.

3:33 If you live next to a coal power plant,

3:35 you're probably, you know, more likely to get asthma.

3:37 Natural gas has things.

3:39 Everything is dangerous.

3:41 So then how do you quantify that danger and those dangers

3:44 from burning fossil fuels to produce electricity to, using nuclear energy?

3:49 There's been some studies, and they really do come down.

3:52 Like, coal has a lot more deaths per, like, terawatt hour power produced.

3:57 Natural gas has more deaths.

4:00 Nuclear power, even including the deaths from Chernobyl,

4:03 much, much lower on the scale.

4:05 I'm gonna just randomly say this, but I'm assuming you get into a car

4:09 and you might probably are more likely to get in an accident Oh, yeah.

4:13 Than a nuclear.

4:14 So That's what I mean.

4:15 This planet is dangerous for us.

4:17 What I wanna ask you about is Microsoft, its vast power needs,

4:19 and this is kind of like the subtitle on the story

4:22 and how it's bringing together two transformative and risky technologies.

4:24 Will I always get nervous when there is such demand and a rush

4:28 in need for something that I wonder

4:30 if people get sloppy in the development process.

4:33 Now I realize everybody's gotta be careful because if there's one problem,

4:36 it will shut down the industry again, probably.

4:39 Certainly The US.

4:41 Question.

4:41 Okay.

4:41 So what what's going on that's risky?

4:46 That's a good question.

4:48 Yeah.

4:48 I mean It it I mean, like you said, Chernobyl was indeed a catastrophe.

4:53 With nuclear I've been to several nuclear plants,

4:56 and everywhere I go, it is just so carefully monitored and screened.

5:02 But let you know, radiation does have, you know,

5:06 the potential to be dangerous, so you have to be careful.

5:09 What go ahead, Tim.

5:10 Well, what I noticed in your piece was

5:13 that there there was this idea that electricity

5:17 demand had kind of peaked because we

5:19 had become more efficient with with energy usage.

5:21 LED bulbs were, you know, actually a part of that pre 2019.

5:25 Mhmm.

5:25 And then this happened.

5:27 And and and by this, I mean, the demand for for AI.

5:31 And I'm wondering about the possibility of overbuilding here because but what

5:35 if it turns out that all of these technologies do become more efficient?

5:39 Because that's the big that's the big bottleneck right now.

5:42 So you have to think that that's what everybody is working toward,

5:45 making this stuff more efficient.

5:46 What happens if we overbuild here and we we just shoot too far?

5:48 We've but then we spent a lot of money.

5:51 Yeah.

5:51 That that's the question that we've been talking about.

5:53 So, I mean, power demand in The US has pretty much been flat for, like,

5:57 twenty something years, and now it's going up.

6:00 Like, the projections are insane on how much power we're going be in the show.

6:06 Projections are projections, though.

6:07 Exactly.

6:07 So, I mean, like, here in, like, second quarter twenty twenty six,

6:12 it's not crazy higher than a year ago,

6:14 but the forecast through 2030 and 2035, they are.

6:18 But, yeah, we AI is brand new.

6:20 This is like a trend I've been following for about two years.

6:23 But if they come up with a way to make it

6:26 more energy efficient all of a sudden, that is a concern.

6:29 And people are, like,

6:31 allocating billions and billions of dollars for power plants.

6:34 And if they overspend, well, we'll see it in the stock price.

6:39 The next generation, though, just to to wrap up here.

6:41 We've talked about the SMRs, the small modular reactors.

6:43 This kind of next generation of technologies.

6:45 Right now, we're doing a lot of, like, reef right?

6:48 Like, existing facilities getting up and running.

6:51 There's three.

6:51 So how long is that, and is that gonna be even more efficient, more safer?

6:56 But what's the timeline?

6:58 The timeline on SMRs is I'm gonna go with, like,

7:01 early twenty thirties, maybe a couple before 2030.

7:04 There's dozens of companies that are working on them.

7:07 There's, like, maybe a half dozen company I

7:09 could name that are making really solid progress, and they're not that far away.

7:13 Right.

7:14 But it's gonna take some time.

7:16 And the first ones are not gonna be cheap.